A 50-100 kWe Gas-cooled Reactor For Use On...

SANDIA REPORT SAND2006-2189 Unlimited Release Printed April 2006

A 50-100 kWe Gas-cooled Reactor For Use On Mars Curtis D. Peters Prepared by Sandia National Laboratories Albuquerque, New Mexico 87185 and Livermore, California 94550 Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear Security Administration under Contract DE-AC04-94AL85000. Approved for public release; further dissemination unlimited.

ii

Issued by Sandia National Laboratories, operated for the United States Department of Energy by Sandia Corporation. NOTICE: This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government, nor any agency thereof, nor any of their employees, nor any of their contractors, subcontractors, or their employees, make any warranty, express or implied, or assume any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represent that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government, any agency thereof, or any of their contractors or subcontractors. The views and opinions expressed herein do not necessarily state or reflect those of the United States Government, any agency thereof, or any of their contractors. Printed in the United States of America. This report has been reproduced directly from the best available copy. Available to DOE and DOE contractors from U.S. Department of Energy Office of Scientific and Technical Information P.O. Box 62 Oak Ridge, TN 37831 Telephone: (865) 576-8401 Facsimile: (865) 576-5728 E-Mail: [email protected] Online ordering: http://www.osti.gov/bridge Available to the public from U.S. Department of Commerce National Technical Information Service 5285 Port Royal Rd. Springfield, VA 22161 Telephone: (800) 553-6847 Facsimile: (703) 605-6900 E-Mail: [email protected] Online order: http://www.ntis.gov/help/ordermethods.asp?loc=7-4-0#online

iii

SAND2006-2189 Unlimited Release Printed April 2006

A 50-100 kWe Gas-cooled Reactor For Use On Mars

Curtis D. Peters (Year Round Intern)

Advanced Nuclear Concepts

Sandia National Laboratories P.O. Box 5800

Albuquerque, New Mexico 87185

Abstract In the space exploration field there is a general consensus that nuclear reactor powered

systems will be extremely desirable for future missions to the outer solar system. Solar

systems suffer from the decreasing intensity of solar radiation and relatively low power

density. Radioisotope Thermoelectric Generators are limited to generating a few

kilowatts electric (kWe). Chemical systems are short-lived due to prodigious fuel use. A

well designed 50-100 kWe nuclear reactor power system would provide sufficient power

for a variety of long term missions. This thesis will present basic work done on a 50-100

kWe reactor power system that has a reasonable lifespan and would function in an

extraterrestrial environment. The system will use a Gas-Cooled Reactor that is directly

coupled to a Closed Brayton Cycle (GCR-CBC) power system. Also included will be

some variations on the primary design and their effects on the characteristics of the

primary design. This thesis also presents a variety of neutronics related calculations, an

examination of the reactor’s thermal characteristics, feasibility for use in an

extraterrestrial environment, and the reactor’s safety characteristics in several accident

scenarios. While there has been past work for space reactors, the challenges introduced

by thin atmospheres like those on Mars have rarely been considered.

iv

ACKNOWLEGEMENTS

The author wishes to thank Ronald J. Lipinski (SNL), Robert Busch (UNM),

Steven A. Wright (SNL), Roger X. Lenard (SNL), and Kendall R. DePriest (SNL) for

providing technical assistance, opportunity and guidance in the creation of this report.

The author also wishes to thank the Sandia National Laboratories (SNL) Student

Internship Program (SIP) for making this report possible.

v

TABLE OF CONTENTS

LIST OF FIGURES ...................................................................................................ix

LIST OF TABLES.....................................................................................................x

1.0 Problem Description and Scope of Design ....................................................1

1.1 Problem Description ..........................................................................1

1.2 Design Scope .....................................................................................3

2.0 Background and Literature Review ...............................................................4

2.1 Gas-Cooled Reactor ...........................................................................4

2.1.1 History of Gas-cooled Reactors .............................................5

2.1.2 How a Gas-cooled Reactor Works.........................................5

2.2 Brayton Cycle ....................................................................................6

2.2.1 Brayton Cycle History ...........................................................7

2.2.2 Brayton Cycle: How It Works ...............................................7

2.3 Fast Spectrum Reactor .......................................................................9

2.4 Fuel ....................................................................................................10

2.5 Pin Type Geometry ............................................................................11

2.6 Space Environment ............................................................................11

3.0 Materials Used in the Design.........................................................................13

3.1 Uranium Nitride .................................................................................13

3.2 Helium/Xenon....................................................................................17

3.3 Niobium 1% Zirconium .....................................................................18

3.4 Hastelloy X ........................................................................................21

3.5 Beryllium Oxide.................................................................................23

vi

3.6 Rhenium.............................................................................................25

4.0 Design Description.........................................................................................27

4.1 Design of Reactor ..............................................................................27

4.2 Gas Flow in Reactor...........................................................................34

4.3 Pressure Vessel Composition.............................................................35

4.4 Core Block Composition....................................................................35

4.5 Reactor Control Methods...................................................................36

5.0 Neutronic Investigation into Core Design Results.........................................38

5.1 Initial Calculations .............................................................................38

5.2 Reactor Multiplication Vs Reflector Position....................................39

5.3 High Temperature Feedback effects ..................................................42

5.4 Accident Scenarios.............................................................................45

5.4.1 Immersion in Water ...............................................................46

5.4.2 Immersion in Wet Sand with Water Flooding .......................46

5.4.3 Dry Sand Burial .....................................................................48

5.5 Alternative Cores ...............................................................................50

5.5.1 Category-III Reactor ..............................................................50

5.5.1.1 Alternative 1 Reactor .................................................52

5.5.1.2 Alternative 2 Reactor .................................................55

5.5.2 Boron Carbide Central Safety Rod ........................................57

5.6 Quality Control Runs .........................................................................62

6.0 Thermal Evaluation of Core and Remainder of System ................................66

6.1 Radiator..............................................................................................67

vii

7.0 Conclusions and Future Work .......................................................................72

7.1 Conclusions........................................................................................72

7.2 Future Work .......................................................................................73

8.0 References......................................................................................................70

viii

LIST OF FIGURES

Figure 1-1. Regions of applicability for different power systems ............................1 Figure 2-1. Closed Brayton Cycle ............................................................................8 Figure 3-1. UN Phase Diagram .................................................................................14 Figure 3-2. UN Mechanical Properties .....................................................................15 Figure 3-3. UN Thermal Properties ..........................................................................16 Figure 3-4. Nb1Zr Operational Domain ...................................................................19 Figure 3-5. Nb1Zr Thermal Properties .....................................................................20 Figure 3-6. Hastelloy X Operational Domain ...........................................................22 Figure 3-7. Hastelloy X Thermal Properties .............................................................23 Figure 3-8. BeO Mechanical Properties ....................................................................24 Figure 3-9. BeO Thermal Properties .........................................................................25 Figure 4-1. XZ Plane Section of Reactor ..................................................................30 Figure 4-2. XY Plane Section of Reactor ................................................................31 Figure 4-3. XZ Plane Section of Individual Fuel Pin ...............................................32 Figure 4-4. XY Plane Section of Fuel Pin ................................................................33 Figure 4-5. Flow Pattern for Reactor ........................................................................34 Figure 4-6. Reflector Movement ...............................................................................36 Figure 4-7. XY Plane Section with Radial Reflectors ..............................................37 Figure 5-1. k-Effective Vs Reflector Gap .................................................................40 Figure 5-2. Reflector Movement ...............................................................................41 Figure 5-3. Flux Profile Vs Position for 3 different reflector settings ......................42 Figure 5-4.Water Immersion Accident Scenario ......................................................47 Figure 5-5. Wet Sand, Water Scenario ....................................................................48 Figure 5-6. Sand Burial Accident Scenario .............................................................49 Figure 5-7. XZ Plane Section of Reactor Alternative 1 ............................................52 Figure 5-8. XY Cross Section of Reactor Alternative 1 ...........................................53 Figure 5-9. Close up of XY plane, Alternative 1 Reactor ........................................53 Figure 5-10. k-Effective Vs Pitch, Alternative 1 Core .............................................54 Figure 5-11. XZ Cross Section of Alternative 2 Reactor ..........................................56 Figure 5-12. XY Cross Section of Alternative 2 Reactor .........................................56 Figure 5-13. Close up of XY Plane, Alternative 2 Reactor ......................................57 Figure 5-14. XY Plane Section of Internal Control Rod Reactor .............................58 Figure 5-15. XZ Plane Section of Internal Control Rod Reactor .............................58 Figure 5-16. Multiplication vs. Run with different RNG seeds ................................63 Figure 5-17. Distribution of Fission Causing Neutrons ............................................64 Figure 6-1. FEPSIM components with State temperatures .......................................66 Figure 6-2. Electric Output vs. Thermal Output .......................................................68 Figure 6-3. Conversion Efficiency Vs Thermal Output ............................................69 Figure 6-4. Specific Power Vs Thermal Output .......................................................70

ix

LIST OF TABLES

Table 3-1 Gas Properties ...........................................................................................18 Table 4-1 Primary Core Parameters...........................................................................28

Table 5-1 Temperature Profile of Reactor .................................................................43 Table 5-2 Temperature Adjusted Dimensions of the Reactor ...................................43 Table 5-3 Temperature Adjusted Material Densities.................................................45 Table 5-4 Accident Scenario Results.........................................................................49 Table 5-5 Reactor Dimensions and Compositions.....................................................60

1

1 Problem Description and Scope of Design

1.1 Problem Description

The need for power sources that can generate more power for longer periods of

time for space applications has been understood for quite some time. Solar powered

systems are good for low power, intermediate-to-long duration missions. Chemical

systems are ideal for high power, short duration missions. Radioisotope Thermoelectric

Generators (RTGs) are good for long duration, low power missions. For long-life, high

power applications nuclear reactors seem to be the best option available, as shown in

Figure 1-1.

Figure 1-1. Regions of applicability for different power systems [LA-7858, 1979]

This plot has a limitation; it is designed for distances from the sun that are comparable to

the earth’s orbit. At distances further from the sun, the region dominated by solar power

2

shrinks drastically. There are other considerations; on the moon, which has a 348 hour

long night, the solar power regime is also much smaller. This plot shows that for systems

requiring more than ten kilowatts of electric power (kWe) for more than a month, nuclear

reactors end up being a favorable power source.

Why are systems that generate more power needed? More power means that more

science instruments can be fielded, a greater amount of data can be transmitted more

quickly, energy intensive applications like electric thrusters become practical, and more

complicated missions are possible. Most current probes fly by their target rather than

orbit it. Probes that use nuclear powered electric drives can place themselves in a stable

orbit around the target and keep examining it. For manned missions, nuclear reactors

generate the kind of power required to sustain life and manufacture fuel for return

missions. Nuclear reactors are a good method for generating this constant and sustained

power.

Currently RTG’s are used to produce power for many probes that are transmitting

data from across the solar system [Furlong, 1999]. However, RTG’s and other systems

are limited in power generation because of the limited amount of available 238 Pu, the

expense of that isotope, and the fuel needed for a multi-kWe system. The weight per

kWe may be unacceptable for some missions. Nuclear fission power systems have the

potential to provide more power than RTG’s. A drawback of reactor systems is that they

are only mass and cost effective for missions requiring more than 10 kWe. Mars and

other planets provide another set of potential challenges. The interaction between the

outer surfaces of a heated reactor and the atmospheres of these planets is far more

complicated than the interaction between the outer surfaces and deep space.

3

1.2 Design Scope

The goal of this thesis is to present a design for a workable gas-cooled, direct Brayton

cycle reactor that could generate 50-100 kWe. The design has not been optimized, but

some sensitivity studies were done. The goals were as follow:

• The reactor shall be a gas-cooled, pin type reactor.

• The power output of the reactor will be roughly 100 kWe.

• There shall be reasonable expectation that the reactor can operate at full power for

10 years.

• The reactor will be safe under credible accident scenarios

• Investigate how to modify the reactor so that it is useable in a Martian

atmosphere.

For purposes of limiting the scope of the thesis, the reactor will use a directly coupled

closed Brayton cycle for power conversion and will use highly enriched Uranium Nitride

fuel. The sections of the thesis will be:

• The primary components of a GCR-CBC reactor

• The properties of the materials used in the reactor.

• A description of the reactor core.

• Neutronics related calculations.

• The radiator and how it affects the power conversion section.

• Conclusions and future work.

4

2 Background and Literature Review

This section covers several topics: the history of the important components for the

reactor, how these components work, and why these components were chosen. The

primary motivation was to design a functional reactor; not to explore all the myriad of

possibilities that exist. This meant that not all the alternatives for any given component

would be examined, just ones that appeared to fulfill the desired goals. As a result

several simplifications were implemented for this thesis to reduce the scope.

The topics that will be examined are:

• What gas-cooled reactors are and how they operate;

• The operation and history of the Brayton Cycle;

• The type of fuel selected.;

• How the fuel is loaded;

• The neutron energy spectrum; and

• The effects of Mars’ atmosphere on the exposed components at the

expected temperatures.

2.1 Gas-Cooled Reactor

A gas-cooled nuclear reactor was chosen due to its simplicity and suitability for the

space environment. There are numerous alternatives for cooling a reactor core, but gas

cooling is one of the simpler methods and most attractive when used with a closed

Brayton cycle (CBC) power conversion system. The use of other reactor coolants would

necessitate the inclusion of a heat exchanger and introduce complicated freeze/thaw

problems, increasing the complexity and the weight of the reactor.

5

2.1.1 History of Gas-cooled Reactors

Gas-cooled reactors have been around since the early days of the nuclear industry.

While none have been flown as space reactors, there is a significant body of work in the

form of test reactors and even commercial power reactors. Notable Gas-cooled

commercial power reactors include the British MAGNOX reactors, Peach Bottom I, and

the Ft Saint Vrain reactor. In addition to power production, gas-cooled reactors have

been suggested as a source of process heat for smelting steel, hydrogen production, and

district heating. These large reactors are significantly different in design from the

proposed reactor. They generally use a different fuel type and their core geometry was

different. TRISO type fuel has been the most common while the British MAGNOX

reactors use natural uranium metal encased in a magnesium oxide cladding. TRISO fuel

is a pellet sized fuel with layers of silicon carbide and pyrolitic carbon applied to the

outside. Most existing gas-cooled reactors use a thermal neutron spectrum and use

graphite as a moderator. The proposed reactor will use pins of highly enriched uranium

nitride (UN) clad in niobium 1% zirconium (Nb1Zr) in a triangular pitch. These pins will

be placed in a prismatic block of Nb1Zr. This limits the comparisons between existing

reactors and the proposed reactor somewhat, but comparisons for the remainder of the

system should still be useful.

2.1.2 How a Gas-cooled Reactor Works

Gas-cooled reactors are fairly simple systems. The working fluid in a gas-cooled

reactor is a single phase gas, which flows across the reactor core [El-Wakil, 1984]. The

gas is heated by the fuel pins and then leaves the reactor core. To remove the heat the gas

is either run through a heat exchanger transferring the heat to some other working fluid

6

or, as in this case, run directly through a turbine to generate power. The advantages of

gas-cooled reactors are:

• The availability of inert gasses.

• Low neutron cross section.

• The ability to run the reactor at high temperatures.

This last factor allows high efficiency power conversion but requires exotic materials to

be practical. As compared to alternative systems, another advantage is the absence of

freeze/thaw issues found in liquid metal based systems. The difficulties of managing two

phase flow in low or zero gravity conditions are also avoided. However, gas coolants are

not as efficient as other coolants in removing heat. While the heat capacity of helium is

superior to that of water or many of the common liquid metals on a per-mole basis, the

density of helium is so low that appreciable heat removal requires large volumes of gas to

be moved across the reactor. This requires careful design for the reactor to avoid large

pressure losses. Pumping large volumes of gas also consumes significant amounts of

power, reducing the amount available for other purposes.

2.2 Brayton Cycle

A Closed Brayton Cycle will be used for the power conversion system. This cycle

has the advantage of being a well understood and robust power conversion cycle.

Extensive testing of similar systems gives confidence in the long-term durability of this

system

7

2.2.1 Brayton Cycle History

Brayton cycle systems have a long history of use and are a well-developed

technology. Commercial jet engines use an open Brayton cycle and provided much of

the initial data for the first adaptation for space power purposes. Gas-fired turbines for

power generation are another application of a Brayton cycle. The development of

Brayton cycle systems for space power applications in the US started in the 1960’s with

NASA’s Brayton Rotating Unit program. The goal of the program was to develop a 10.5

kWe Brayton system that could be used with both solar and radioisotope systems [Davis,

1972]. Four different units were tested with a combined 40,000 hours of operation. This

program set the stage for a succession of smaller and larger Brayton systems. During the

1980’s there was another revival of the program with the intent to deploy a solar-heated

Brayton System [NASA, 1993]. Thus, a long standing well tested program gives reason

to believe that Brayton Cycle systems are a viable long-life power conversion system for

space applications.

2.2.2 Brayton Cycle: How It Works

There are two primary variants of the Brayton cycle: the open cycle and the

closed cycle [El-Wakil, 1984]. In an open cycle system, the coolant is drawn in from the

outside environment, heated, run through a turbine, and discharged back to the outside

atmosphere. In a closed cycle, the gas is in a closed loop and used repeatedly. There are

two variants of closed Brayton cycles: direct and indirect. In the direct system, the heat

source is directly coupled to the gas flow system while in an indirect system, the coolant

passes through an intermediary heat exchanger. A block diagram of a direct closed

Brayton cycle is shown below:

8

Figure 2-1 Closed Brayton Cycle [El-Wakil, 1984]

The working fluid, a single phase gas, is heated in some manner. For gas turbines used as

peaking power systems, this means combusting a gas. For reactor-based systems, the gas

is run through the reactor core. This gas is then run through a turbine, converting thermal

energy into work. The gas is then either vented to the atmosphere (open system) or run

through a heat exchanger or radiator to lower its temperature, compressed, and then fed

back into the heat source (closed system).

Brayton conversion systems have advantages and disadvantages. They are more

efficient than most static power conversion systems (e.g., thermoelectric or thermionic

based systems), and they are more durable and simpler than the other dynamic power

conversion systems. However, Brayton cycles do require higher temperatures to achieve

the same efficiency as other dynamic power conversion systems. The energy density of

the working fluid is low compared to the other dynamic power conversion systems.

9

Single point failure can be a problem with Brayton systems. A leak in the reactor can

easily lead to all the coolant being lost.

2.3 Fast Spectrum Reactor

The decision to focus on a fast spectrum reactor was largely a result of the desire

for a high temperature gas outlet. The higher temperature of the gas increases the

efficiency of the Brayton cycle. Thermal systems tend to be larger than fast systems.

Thermal spectrum systems require moderators, and efficient moderators tend to be

hydrogen-based and most are not compatible with high temperatures. Two possible high

temperature moderators are yttrium hydride and graphite. Yttrium hydride is a hydrogen–

based, high temperature moderator that is currently being examined. It has some serious

unknowns associated with it including the long-term effects of being at the desired

elevated temperatures in a high-radiation environment and the temperature feedback

coefficient. Graphite is much less efficient at moderating neutrons and would likely

increase the size of the reactor significantly. This makes fast spectrum reactors favorable

if high temperature outlet gases are desired.

Fast Spectrum reactors use high energy neutrons (>100 keV) to initiate fission.

There are a variety of complications associated with this. At these higher energies,

fission cross sections drop from the multi-hundred barn range into the low single digit

range, reducing the reaction rate significantly. Thus for the same power output, a fast

reactor needs to have a much higher neutron flux than a thermal reactor. This tends to

accelerate radiation induced damage to other components of the reactor. On the other

hand, the relative magnitudes of the cross sections of materials mean that most materials

are relatively transparent to fast neutrons. Boron, rhenium, and a variety of other

10

materials that are poisons in the thermal spectrum have a far smaller impact on fast

neutrons. The method for controlling the reactor is external reflectors. The position of

the reflectors is used to adjust the leakage of the system. Other than a slight diversion into

what a thermal spectrum reactor would look like, the vast majority of this project will be

done for fast cores.

2.4 Fuel

The fuel type selected for the core is highly enriched UN because UN has been

tested to a greater degree than uranium carbide, which is theorized to have similar

characteristics. The thermal properties of UN are favorable as compared to UO2; the only

fuel type that has been tested to a greater degree. UN is a better choice than the TRISO

type fuel typically used in gas-cooled reactors because of the low relative uranium

density (7-8% of UN) of TRISO. One of the primary advantages of TRISO type fuels is

the high achievable burnup. As high burnup of the fuel is not one of the design goals,

TRISO’s primary advantage is negated. TRISO’s reduced uranium density would

necessitate an unacceptable increase in core size. Testing has shown that UN is a viable

material at the expected burnup for this core.

The fuel in the reactor will be 93% enriched UN. The decision to use Highly

Enriched Uranium (HEU) for fuel is partially a result of the lifespan related

considerations. Ten years at 200- 400 kW thermal power would consume 0.8-1.6 kg of

235Uranium. With the proposed highly enriched system this is only 0.86% of the fuel,

which is a small burnup. The reactivity swing associated with this burnup was

determined to be 0.006 (~1$), and sufficient extra fuel needs to be in the core to

compensate for this reactivity swing. HEU does increase the security related costs for the

11

reactor. One minor alternative designed to mitigate this cost for a nuclear heated ground

test facility is examined in the alternative cores section.

2.5 Pin Type Geometry

The fuel pins used in the core are a new type for GCR’s. The vast majority of

GCR reactors use a TRISO type fuel embedded in a graphite matrix and are thermal

neutron spectrum reactors. TRISO fuels are small UC spheres coated with layers of

silicon carbide and pyrolitic carbon. While this results in fuels that can be used to

extremely high burnup, the uranium density of the fuel is low. A coolant hole is then

drilled through the blocks of fuel and these blocks are put in a prismatic array.

The proposed reactor uses an array of UN pellet filled pins in a triangular pitch.

These pins are clad in a layer of rhenium and Nb1Zr. The use of rhenium limits the

problem of nitrogen attack on Nb1Zr. It also has positive accident safety characteristics.

These pins are similar to those developed for the SP-100 program though they are used in

a different reactor setup.

2.6 Space Environment

This reactor is intended for use on Mars. Unlike many other extraterrestrial

environments, Mars has an atmosphere. Specifically the atmosphere is roughly 4-7

millibar of pressure and composed of 95% carbon dioxide, 2.7% nitrogen, 1.5% argon,

0.15% oxygen, and 0.15% water (Keifer, 1992). This presents a host of complications

for materials choices in the reactor. The pressure vessel of the reactor will be exposed to

this environment while at an elevated temperature. While this temperature is lower than

what one would expect for a liquid metal or rankine cycle reactor, it is high enough to

12

accelerate reactions. The interaction between the environment and the outer surfaces of

the reactor is a problem that will not be examined in this work. Oxidation and other

reactions will cause problems for reactors and must be considered. For example, at the

likely temperature for the pressure vessel (876 K), Nb1Zr absorbs oxygen from the

atmosphere readily and becomes brittle [DiStefano, 1990]. The material used on the

exterior of the reactor needs to be resilient enough to last the expected 10 years in the

given environment. This tends to eliminate many of the high temperature refractory

alloys which have oxidation problems.

13

3 Materials Used in the Design

A variety of relevant properties at the estimated temperatures for the given material

will also be examined. There are three general classes of materials in the reactor: Metals,

Ceramics, and Gasses.

Metals

For all the metals several properties will be examined. The mechanical properties

are Yield Strength (YS), Ultimate Tensile Strength (UTS), and Creep Strength Vs

Temperature. The thermal properties that will be examined are thermal conductivity and

linear expansion.

Ceramics

For the Ceramic Materials, the mechanical property that will be examined is the

Modulus of Elasticity. The thermal properties that will be examined are thermal

conductivity and linear expansion.

Gasses

The helium/xenon coolant will be compared relative to some other common gasses.

The properties that will be examined are viscosity, thermal conductivity, specific heat and

density.

3.1 Uranium Nitride

UN is a ceramic mixture of uranium and nitrogen. It has a melting point in the

range of 3100 K, well above the likely peak fuel temperatures (1300 K) [Tagawa, 1974].

The maximum theoretical density is 14.32 g/cc [Johnson, 1976]. The core will use a 1-

to-1 atomic ratio mixture. A phase diagram for UN can be found in Figure 3-1.

14

Figure 3-1 UN Phase Diagram [Tagawa, 1974]

Peak Fuel Temp (1030 C)

15

Mechanical Properties

UN Mechanical Properties

195

200

205

210

215

220

225

700 900 1100 1300 1500 1700 1900

Temp (K)

Stre

ss (G

Pa)

Modulus of Elasticity

Figure 3-2 UN Mechanical Properties

Thermal Properties

There are a variety of advantages to using uranium nitride (UN) over the other

fuel types, and a handful of drawbacks that need to be considered. Uranium nitride has a

much higher thermal conductivity than uranium dioxide, resulting in a flatter temperature

profile across the fuel pin. The same is likely true for uranium carbide, but less testing

has been done on uranium carbide than for UN. Figure 3-3 shows the desired thermal

properties of UN [Touloukian, 1979]. At the operational temperature, 1300 K, the

thermal conductivity is 28.5 W/m-K and the linear expansion is roughly 1%. It is

important for the different components to expand at similar rates to minimize the extra

stresses.

Expected Temperature

16

UN Thermal Properties

0

5

10

15

20

25

30

35

800 1000 1200 1400 1600

Temp (K)

Ther

mal

Con

duct

ivity

(W

/m-k

)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

Line

ar E

xpan

sion

(%)

Thermal ConductivityLinear Expansion (%)

Figure 3-3 UN Thermal Properties

Additional Considerations

There are contradictory statements on fission fragment retention from different

sources. There is a general statement that UN retains fission fragments better than UO2.

In personal discussions others have indicated that the oxygen released due to fission for

UO2 systems has a tendency to bind with several types of fission fragments into an oxide

that does not migrate [Lenard, Roger and Lipinski, Ronald, personal communication,

August 2005]. While the latter does not necessarily invalidate the general sweep of the

former statement, it does highlight the difficulties in making sweeping generalizations.

In this design, the UN fuel will be located in the drum of the core, clad in a layer of

rhenium and niobium 1% zirconium. This rhenium layer is required because of

incompatibilities between nitrogen and niobium 1% zirconium. The nitrogen out-gassing


17

and the fission gasses that are not bound by the nitrogen aggressively attack Nb1Zr

cladding, leading to breaches into the clad. There is evidence that this problem is

widespread across a variety of materials. The effects of irradiation on UN are still not

completely understood.

3.2 Helium/Xenon

The coolant chosen for the reactor will be a mixture of helium and xenon. There

are two primary reasons for choosing a mixture: the limitations related to machining

components and the favorable thermal characteristics of He/Xe mixtures. The vast

majority of the parameters that go into determining the turbine and compressor wheel

size are fixed by other limitations. If a low molecular weight gas is used, the size of the

turbine wheel will end up being small. The tolerances between the wheel blades and the

cowling are the same regardless of the wheels size so for a smaller wheel the magnitude

of the losses goes up drastically. [Steven Wright, personal communication, March 20,

2006] One method of avoiding these losses is to use a heavier gas. Heavier gasses tend to

have lower heat capacities. Mixtures of helium and xenon have favorable thermal

characteristics when compared to argon and neon. A mixture of helium and xenon can

have the same average molecular weight as argon or neon but have better heat transfer

characteristics than either one. A mixture of helium and xenon with an average

molecular weight of around 40 g/mol seems to be a very common choice [Angelo, 1985].

Table 3-1 shows properties of helium, xenon, the proposed mixture, and other gases for

comparison. The properties for all the gases except for xenon were taken at 1000 K and 2

MPa. Xenon was taken at 800K due to limitations of the reference.

18

Table 3-1 Gas Properties [1] [2]

Helium

[1]

Xenon

[1]

He/Xe

[2]

(70/30) N2 [1]

CO2

[1]

Density (kg/m^3) 0.96 39.46 7.5 6.69 10.5

Specific Heat (Cp)

(J/mol-k) 20.78 21.151 20.52 32.75 54.5

Gamma 1.67 1.69 1.676 1.34 1.18

Viscosity (μPa-s) 46.17 54.908 65.2 41.68 41.31

Thermal conductivity

(W/m-k) 0.3617 0.01319 0.15446 0.0661 0.0708

[1] From NIST Chemistry Webbook, http://webbook.nist.gov/chemistry/

[2] From Lipinski, 20002

This table shows that helium has the highest thermal conductivity with the helium/xenon

mixture coming in second.

3.3 Niobium 1% Zirconium

Niobium 1% zirconium is an alloy of 99% niobium and 1% zirconium with a

melting point of Nb1Zr is 2673 K and a density of 8.64 g/cc at room temperature

[Summary, 1965]. Nb1Zr is a robust alloy that retains its strength up to high

temperatures. The nominal temperature limit for the alloy is 1350 K. The expected peak

temperature for Nb1Zr as a cladding is 1300 K, below its nominal limit.

19


Figure 3-4 shows the operational domain of Nb1Zr with vertical line showing the

expected operating temperature 1300 K.

Nb1Zr Operation Domain

1

10

100

1000

400 600 800 1000 1200 1400 1600

Temp (K)

Stre

ngth

(MP

a)

1/3 UTS.2% Offset YS1% Creep, 10 5̂ Hours1% Creep, 10 4̂ hours

Figure 3-4 Nb1Zr Operational Domain

This chart represents minimum values expected for Nb1Zr [Summary, 1965] [Watson,

1965]. Several other sources listed higher values for both the yield strength and tensile

strength; some by as much as 30% [Niobium, 2006]. This plot shows that at the expected

temperatures, the maximum yield strength is roughly 68 MPa and the 1/3 tensile strength

limit is 52 MPa. In any event, at the desired temperatures, neither of these properties is

the limiting factor; creep strength is. The 1% creep strengths of Nb1Zr in a liquid

lithium environment are also shown in Figure 3-4. While the Nb1Zr will be in a He/Xe

environment, it was the best data available. The 1% creep limit at 105 hours (roughly 11

years of continuous operation) shows the stress limit on the niobium 1% zirconium


20

[Horak, 1985] is about 7.38 MPa. Of the possible limiting stresses for the metal this is

the lowest at 1300 K and thus becomes the limiting property for Niobium 1% Zirconium.

Thermal Properties

The thermal conductivity of Nb1Zr at 1300 K is 69 W/m-K [Touloukian, 1970]. The

linear expansion of Nb1Zr at 1300 K is 1% [Touloukian, 1975]. Figure 3-5 shows linear

expansion and thermal conductivity as a function of temperature.

Nb1Zr Thermal Properties

50

55

60

65

70

75

80

500 700 900 1100 1300 1500

Temperature (K)

Ther

mal

Con

duct

ivity

(W

/(M-K

)

0.000

0.100

0.200

0.300

0.400

0.500

0.600

0.700

0.800

0.900

1.000

Line

ar E

xpan

sion

(%)


Figure 3-5 Nb1Zr Thermal Properties

Additional Considerations

There are some limitations to Nb1Zr based alloys. At high temperatures Niobium

oxidizes rapidly in atmospheres containing oxygen, resulting in significant embrittlement

[DiStefano, 1990]. This restricts the use of Nb1Zr when in contact with most

atmospheres. These limitations preclude the use of Nb1Zr on the outside surfaces of the


21

reactor. The strength of Nb1Zr is also limited. There has been little indication that such a

breakthrough has been made to date.

Alternatives worth Pursuing

There is a slight variant called PWC-11 that introduces 0.1% carbon into the Nb1Zr alloy.

This increases the creep strength properties by 30-40% while not changing the thermal

properties. It would probably merit further investigation as a material choice.

3.4 Hastelloy X

Hastelloy X (HastX) is a nickel-based superalloy used in a variety of applications

[Brown, 1992]. Its composition is 49% nickel, 22% chromium, 18% iron, 9%

molybdenum, 1.5% cobalt and 0.5% tungsten. The melting point of HastX is about 1530

K, and it has a density of 8.22 g/cc. This material has been suggested as a possible

reactor material for a variety of reasons. HastX is a material with decent high

temperature characteristics. Hastelloy is also noted for excellent corrosion, oxidation and

carburization resistance at the desired temperatures. Finally, Hastelloy-X is a commonly

used metal whose properties are well understood. The expected peak temperature of

Hastelloy-X is roughly 875 K when used for the pressure vessel of the reactor.

22


Figure 3-6 shows the Yield Strength, 1/3 UTS, and 1% creep at 10000 hours

(approximately 1 year).

Hast-X Operation Domain

1

10

100

1000

600 700 800 900 1000 1100 1200

Temp (K)

Stre

ngth

(MP

a)

1/3 UTS.2% offset YS1% Creep, 10 4̂ Hours

Figure 3-6 Hastelloy X Operational Domain

There are some notable limitations to this data. The 1% creep data for 10^5 hours was

not available. The 10^4 hours creep data did not extend to the desired temperatures. The

1/3 UTS is roughly 195 MPa and the 0.2% offset yield strength was 245 MPa at 860 K.

In all likelihood, all three limits are close to the same value at the expected operation

temperature. The creep strength is constraint most likely to be exceeded

Thermal Properties

The thermal conductivity of HastX is 20.8 W/m-K at 860K and the linear expansion is

0.8%, as shown in Figure 3-7.


23

Hastelloy Thermal Properties

0

5

10

15

20

25

30

600 700 800 900 1000 1100 1200

Temperature (K)

Ther

mal

Co

nduc

tivity

(W

/m-k

)

00.20.40.60.811.21.41.61.8

Line

ar

Expa

nsio

n (%

)


Figure 3-7 HastelloyX Thermal Properties

3.5 Beryllium Oxide

BeO is a ceramic with a 1-to-1 atom ratio of beryllium to oxygen. It is primarily

used in the nuclear field as a neutron reflector. The oxygen component results in BeO

being an extremely good scattering material as it has a very high scattering to absorption

cross section ratio. In addition to being a reflector, there is an n-2n reaction in beryllium

that increases neutron numbers, increasing this material’s attractiveness. BeO will be

located in two different regions of the core with drastically different temperatures. In the

individual fuel pins, axial reflectors of 5 cm BeO at the pin ends will be at around 1300

K. The radial reflectors will be at some temperature between 860 K and the

environmental temperature.


24


Modulus Of Elasticity

0

50

100

150

200

250

300

350

400 600 800 1000 1200 1400 1600

Temp (K)

Mod

ulus

(GP

a)

Modulus Of Elasticity

Figure 3-8 BeO Mechanical Properties

Thermal Properties

BeO has a thermal conductivity of 0.635 W/cm-K at 860 K and 0.28 W/cm-K at 1300 K.

The linear expansion is0 .45% at 860 K and0 .912% at 1300 K

Expected Temperature (860 K)


25

BeO Thermal Properties

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

400 600 800 1000 1200 1400 1600

Temp (K)

Ther

mal

Con

duct

ivity

(W/c

m-K

)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Line

ar E

xpan

sion

(%)

Thermal Conductivity (W/cm-K)Linear Expansion (%)

Figure 3-9 BeO Thermal Properties

3.6 Rhenium

Rhenium is a super dense refractory metal that will be used in a non-structural

capacity in the reactor. Its melting point is 3459 K, and its density is 21 g/cc. There are

a variety of characteristics of rhenium that make its use a challenge. There seems to be a

great deal of variability in the mechanical properties of rhenium, depending greatly on the

methods used to prepare the sample [Biaglow, 1995]. All of these things complicate any

attempt to use it in the reactor, but the amount of rhenium in the core is limited and not

used in a structural capacity. Rhenium is used as a liner 0.062 cm thick between the

cladding and the fuel. Thermally speaking, there is an almost negligible drop of

temperature across the liner. Rhenium is not used as a structural material so its



26

mechanical properties are not that important. The thickness of the rhenium layer is so thin

that its thermal properties have a minimal impact on the core.

Attractive Features

One of its attractive features of rhenium is that it is a spectral shift absorber (SSA), which

means that it has a low relative absorption cross section for fast neutrons; while in the

thermal spectrum its absorption cross section increases dramatically. This has safety

applications for the reactor design in accident scenarios. Rhenium has an absorption

cross section of in the fast spectrum, however the magnitude of the difference between

the absorption cross section and the fast fission cross section of 235U is low compared to

the difference at a thermal spectrum. It also provides a barrier that protects Niobium 1%

Zirconium from nitrogen attack and damage caused by other fission products that outgas

from the fuel. Most of the other SSA materials have a relatively low melting point,

making them less attractive.

27

4 Design Description

4.1 Design of Reactor

The original parameters of the reactor design were derived from an internal study

done by Ron Lipinski, Dan Dorsey et al. for an earlier project. The general goal of that

project was to explore the region in which a gas-cooled pin type reactor was practical.

They did a number of runs where the radius of the fuel, thickness of the rhenium liner,

thickness of the cladding, and number of fuel pins were varied. However, the runs were

limited calculations at room temperature. The dry sand accident case was not examined,

and there was no attempt at optimizing the designs. Parametric scans showed the

multiplication factor for these cases, and the results provided a starting point for the

reactor dimensions. The effect of the reflector position on the reactor multiplication

factor was not examined either. Unfortunately this work was not published but the results

were available to the author.

What this thesis will do is:

• Examine the effect of the reflector on the reactor

• Determine the temperature effects on the core

• Optimize the primary core for the three accident scenarios

• Examine alternate cores for

o Reduced testing costs (Cat-III cores)

o Enhanced Launch Safety (Internal Control Rod)

• analyze Quality Control runs to better understand the inner workings of

the reactor

28

• Look at the effects of the radiator on the Brayton cycle efficiency and

system mass.

One of the initial considerations was the desire to keep the UN fuel pellets similar

to those used on the SP-100 project as it was known that fuel could be manufactured at

those dimensions. This fixed the number of fuel pins and other parameters. The reactor

is controlled with external reflectors that move axially to open up a gap between the

upper and lower reflectors, increasing the leakage of neutrons from the center of the

reactor. The fuel is cooled by a 70/30 molar fraction Helium/Xenon gas mix. Table 4-1

shows a variety of parameters for the reactor:

Table 4-1: Primary Core Parameters

Core Primary Design Gas Properties Coolant He/Xe He fraction 70/30 Reactor Core, Vessel, Reflectors Type Pin-matrix Reactor Material Properties Fuel material UN Clad material Nb1%Zr Clad liner material Re Wire wrap material Re Moderator material N/A Matrix (core block) material Nb1%Zr Pressure vessel material HastX Radial reflector material BeO Axial reflector material BeO Lower grid material Nb1%Zr Upper grid material Nb1%Zr Coolant material He/Xe Fissile material U-235 Fuel enrichment 0.9315 Reactor Radial Dimensions Radius of the fuel (UN only) (m) 4.54E-03

29

Thickness of the fuel gap (m) 6E-05 Thickness of the liner (m) 6.2E-04 Thickness of liner-clad gap (m) 4E-05 Thickness of the clad (m) 4.4E-04 Thickness of coolant channel (m) 1.65E-03 Width to thickness ratio of rectangular wire 1 Pitch of the wire wrap (m) 0.2 Number of wires wraps per pin 2 Thickness of matrix between pins (m) 7.8E-4 Number of fuel pins 451 Thickness added to circle to get matrix radius (m) 0.016 Thickness of matrix insulation (He/Xe) (m) 0.0002 Thickness of the matrix & dome baffle (m) 1E-06 Thickness of the downcomer channel (m) 0.008 Thickness of the pressure vessel (m) 0.003 Thickness of the reflector gap (m) 0.001 Thickness of the radial reflector w/o clad (m) 0.15 Reactor Axial Dimensions Length of the active fuel in the reactor (m) 0.52 Length of the axial fuel plenum (m) 0.03 Length of upper axial reflector (m) 0.05 Length of lower axial reflector (m) 0.05 Length of each end cap (m) 0.005 Length of the lower grid plate & pin flow orifices (m) N/A Length of the outlet (upper) plenum (m) 0.005 Length of the inlet (lower )plenum (m) 0.005 Other Properties Axial peak-to-avg ratio in core 1.2

30

Radial peak-to-avg pin power 1.2 SiO2 fraction in wet sand 0.64 K-Effective of Core 1.037 ±0.001

The following are graphics showing the reactor as analyzed. Figure 4-1 shows an axial

cross section of the reactor with the reflectors partially open.

Figure 4-1 XZ Plane Section of Reactor

Gas/Environment He/Xe UN Fuel Pins Pressure Vessel BeO Reflector

31

Figure 4-2 XY Plane Section of Reactor

Figure 4-2 shows an XY plane of the reactor core without the radial reflectors. Moving

in radially we see:

• The pressure vessel of the reactor

• The downcomer that the cold inlet gas flows through

• The core block with the fuel pins suspended in it

Flowing coolant through the downcomer keeps the surface of the reactor at a lower

temperature. The diameter of the barrel of the core without the reflectors is 19.5 cm.

Figure 4-3 shows an axial slice of a single fuel pin with a close-up of the end cap section.

Gas/Environment Nb1%Zr Core Block He/Xe UN Fuel Pins Pressure Vessel Downcomer

32

Figure 4-3 XZ Plane Section of Individual Fuel Pin

The fuel consists of a series of UN pellets that are total of 52 cm in length and 0.91 cm in

diameter. These are capped by 5 cm of BeO to act as an axial reflector. There is a gas

plenum on the top of the pin to allow for fission product gas build up. A layer of rhenium

acts as a liner for the fuel pin to prevent interaction between the Nb1Zr cladding and the

UN pellets and also acts as a thermal neutron absorber in accident cases as discussed

later. The Nb1Zr end caps appear large in this Figure but their purpose is to simulate

other components like the connectors that attach the fuel pin to a grid plate.

Nb1%Zr End Caps Gas Plenum Re/Nb1%Zr Cladding BeO Axial Reflector UN Fuel He/Xe Gas Flow Gap

33

Figure 4-4 XY Plane Section of Fuel Pin

The XY plane section of a pin shows a close-up of the radial layout of a single pin. The

innermost volume of the fuel pin is UN. Then there is a small gas gap followed by a

layer of Rhenium. Next is the Nb1Zr cladding and then the gas flow channel. Finally

there is the matrix that the fuel pins are suspended in. A rhenium wire wrap is used to

keep the fuel pins centered in the flow channels.

Gas flow channel Rhenium Liner Rhenium Wire wrap Nb-1% Zr Cladding UN Fuel Nb-1%Zr/Hast-X Matrix block

34

4.2 Gas Flow in Reactor

Figure 4-5 Flow Pattern for Reactor

One of the more interesting characteristics of the reactor is how the flow pattern is used

to cool the pressure boundaries of the reactor. Figure 4-5 of a similar reactor and used

only for illustration purposes [Brown, N., 2006]. The exterior of the reactor is made out

of HastX, which has a much lower tolerance for high temperatures than the Nb1Zr that

makes up the bulk of the reactor, thus the need for cooling. The top hemisphere of the

reactor is actually two separate shells. Cold gas flows in through the outer shell (1). It

1

2

3

4

5

35

then passes through an annulus on the outside edge of the fuel pin matrix(2). This keeps

cool inlet gas in contact with the outer pressure vessel reducing its temperature to 850 K

from the 960K of the outer surface of the fuel pin block. In the bottom hemisphere, it

then loops around (3) and is forced up through the coolant channels around the fuel pins

(4). Finally the gas exits the top of the reactor on its way to the turbine (5).

4.3 Pressure Vessel Composition: Hastelloy-X

As a result of the desire for a pressure vessel material compatible with the Martian

environment at the desired temperatures Hastx was chosen for the pressure vessel. HastX

is known for being extremely insensitive to corrosion and carburization [Brown, W.,

1992]. Theoretically, the relatively small neutronic penalty associated with Hastelloy

means that any comparably transparent material (from a neutronic standpoint) could be

substituted, which may be necessary if HastX and Nb1Zr are incompatible.

4.4 Core Block Composition

Initially a decision was made to make the core block (the matrix of material that

the fuel pins sit in) out of either HastX or Nb1Zr. There is little neutronic difference

between the two materials. The deciding factor between the two is the peak temperature

in the core. If a lower temperature (and efficiency) is acceptable, then HastX is probably

the better choice. Niobium 1% Zirconium is the higher-temperature alternative. To

achieve hicher efficiency the core block of the reactor will be made out of Nb1Zr. At 100

kWe the temperature in the block is high enough that the long term creep strength of

HastX is a concern.

36

4.5 Reactor Control Methods

The reactor will be controlled with external reflectors that slide axially up and down

along the reactor creating a gap adjacent to the active fuel length of the reactor as shown

in Figure 4-6.

Figure 4-6 Reflector Movement

Movement of the reflector elements increases or decreases the leakage of the reactor,

allowing for control of the neutron population. The reflector is made with two seperate

components: the moveable reflector and the fixed reflector. The upper one-third of the

reflector is fixed in place forming an annulus around the upper part of the reactor. The

lower moveable section provides variable leakage for control of the reactor core. The gap

where the reflectors open up is in a region of relatively high flux so small changes in

reflector position have a larger impact on the leakage. The moveable reflector consists of

eight separate segments to minimize the risk of mission failure should a moveable

reflector fails. The eight segments of the reflector are shown in Figure 4-7.

37

Figure 4-7 XY Plane Section with Radial Reflectors

Separating the reflector into segments is a good idea because during launch the reflectors

will be ‘open’ and having one stick would reduce the useable life of the reactor by

limiting the available reactivity swing. If a reflector were to stick in the ‘closed’ position

the problems are lessened; there is still a significant amount of reactivity swing available

from moving the remaining blocks. There is a chance that the reactor will initially have a

‘safety block’ along the centerline of the core to ensure sub-criticality while the system is

being launched, but control of the reactor after launch will be done solely with external

radial reflectors.

38

5 Neutronic Investigation into Core Design Results

To examine its viability, a variety of parametric scans were performed using MCNP-

5 [LANL, 2003] on a model of the proposed core. The core was designed with sufficient

reactivity to ensure 10 years of full-power operation even when including such negative

feedback coefficients as temperature, temperature based expansion, and burnup. The

neutronics runs done include:

• A series of studies to ensure that the external reflectors could provide enough

of a reactivity swing to control the reactor over its lifespan. Additionally, the

flux profile at the surface of the reactor at different reflector settings was

examined.

• A set of calculations to estimate the negative feedback associated with the

thermal expansion of the reactor and the Doppler broadening associated with

the higher component temperatures.

• A series of runs that simulated some of the credible worst case accident

scenarios.

• Two variations on the main cores to fulfill slightly different requirements.

• Other runs to ensure the accuracy of the results and confirm that the core was

operating as expected. These include runs using different seeds for the

random number generator to ensure that the sampling of the core for the

neutrons is sufficient and runs examining the spectrum of the neutrons

causing fission in the core in accident cases.

• Additionally, a study to determine the effect of the Nb1Zr cladding on the

neutronics of the system.

39

5.1 Initial Calculations

This section will cover some simple calculations related to the reactor. The reactor

had a cold, beginning of life, neutron multiplication factor of 1.037 ±0.001, which

corresponds to an excess positive reactivity of $5.7 based on a delayed neutron fraction

of 0.0065. The burnup for the reactor was determined using a fairly simple set of

equations. The consumption over 10 years at a power level of 200 kWth was 0.8 kgs of

235U and at 400 kWth, 1.6 kgs of 235U would be consumed. Given that the fuel loading is

186 kgs of 235U, the burnup is ~0.86% for the upper end of the uranium consumed. This

burnup results in a loss of $1 reactivity.

5.2 Reactor Multiplication Vs Reflector Position

The primary goal of this study was to ensure that there was sufficient excess

reactivity in the neutron multiplication factor to keep the reactor critical for the 10 year

lifespan while ensuring that the reactor would be subcritical during major accident

scenarios. The position of the reflector can be used to set the multiplication factor of the

reactor. Burnup in the reactor causes a proliferation of additional materials to absorb

neutrons and reduces the density of fissile materials, lowering the neutron multiplication

of the reactor. This can be offset by closing the reflector, which decreases the neutron

leakage of the system. This is shown in Figure 5-1.

40

k-effective Vs Reflector Gap

0.84

0.86

0.88

0.9

0.92

0.94

0.96

0.98

1

1.02

1.04

1.06

0 5 10 15 20 25 30 35

Reflector Gap (cm)

k-ef

fect

ive

Keff vs ReflectorPositionNo Sliding Reflector

Figure 5-1 k-Effective Vs Reflector Gap

The change in neutron multiplication vs reflector position is nearly linear for an extended

region. Beyond 22 cm it starts curving, asymptotically approaching the neutron

multiplication factor of the core without reflectors. The combined worth of the moveable

reflectors is roughly $26.5 from full open to full closed position. Figure 5-2 shows how

the reflectors open up on the core and where the centerline of the reactor is. It also

indicates the length of the reflectors.

41

Figure 5-2 Reflector Movement Figure 5-3 shows the results of a series of neutron flux tallies in the pressure vessel. The

pressure vessel was segmented into eight different axial pieces of equal size. The plot

shows that the axial flux of the reactor is symmetric when the reflectors are closed. As

the reflectors open up, the gap causes a distortion in the flux especially adjacent to the

gap. This is in line with what would be expected for the reactor.

Fixed Upper Reflector (20 cm tall) Gap Centerline Lower Reflector (42 cm tall)

42

-20

-15

-10

-5

0

5

10

15

20

1.00E-04 2.00E-04 3.00E-04 4.00E-04 5.00E-04 6.00E-04

Flux

Cm

off

Cen

terli

ne

0 cm Gap10 cm Gap20 cm Gap

Figure 5-3 Flux Profile Vs Position for 3 different reflector positions

5.3 High Temperature Feedback effects:

Most of the calculations run use room temperature cross sections and room temperature

dimensions as this is standard practice for determining the rough dimensions and

geometry of the core. These runs neglect the effect on cross sections resulting from the

increasing temperature, and they also ignore the effects of elevated temperature on the

dimensions of the reactors. The cumulative effect of both is investigated in this section.

The temperatures used in this section for determining the expansion of reactor

components came from an unpublished code, Fission Electric Power SIMulation

(FEPSIM) [Lipinski, 2002]. These temperatures are shown in Table 5-1.

43

Table 5-1 Temperature Profile of Reactor

Temperature (K)

*Fuel Peak 1305*BeO Axial Peak 1300*Rhenium 1291.6*Clad 1291.6Gas Inlet 850Gas Avg 1000*Gas Peak 1144*Matrix Peak 1217.8Matrix Edge 960*PV 860*Beo Radial 860* indicates temperatures used

The assumption is that once the feedback associated with temperature is known the

negative feedback coefficient for the reactor can be estimated. This will determine if

sufficient excess reactivity is present for the desired mission lifetime. The dimensions of

the reactor when heated to the operating temperatures of 1300 K for peak fuel

temperature and 860 K in the pressure vessel are shown in Table 5-2:

Table 5-2 Temperature Adjusted Dimensions of the Reactor

Old

Radii/Height

New

Radii/Height

Change in

Thickness

UN/BeO Radial 0.454 0.4591 0.00510

GasGap1 0.46 0.465 -0.00010

Rhenium 0.522 0.5272 0.00020

GasGap2 0.526 0.5311 -0.00010

44

Cladding 0.57 0.5755 0.00040

Flow 0.735 0.7392 -0.00130

Cell 0.774 0.778 N/A

UN Height 52 52.583 0.58300

BeO Height 5 5.058 0.05800

Gas Plenum 3 3 0.0

Cap 0.5 0.5 0.0

Height 66 66.699 0.69900

Core Radius 19.65 19.77 0.0

One of the conditions imposed was that the cross sectional area of the gas flow gaps

would not increase. This was a simplification and as the outer radius of the fuel moved

outward to preserve area, the outer radius of the gas gaps expanded slower, resulting in a

net decrease in thickness. Note that the change in thickness expressed in the third column

does not take into account movement outward caused by expansion of interior

components. The cladding is an excellent example of this; its outer radius expands from

0.57 cm to 0.5755 cm, a 0.0055 cm movement. But the bulk of that expansion occurred

in the fuel itself and only a tiny fraction (.0003) was a result of expansion of the Nb1Zr

cladding. Additionally, the expansion of the hexagonal cells that the fuel pins are

suspended in was a more complicated calculation; the ‘radius’ listed is really the ½ pitch

between the pins. Slight increases in the thickness at that point results in larger increases

in volume than would be the case for cylinders. The expansion of the core radius was a

matter of taking the new size of the unit cell and applying it across the centerline of the

45

core. Since there were 23 pins across the core on the center plane, the radius was

increased by that expansion amount. The densities of the materials in the core were also

altered to compensate for their increased volume. This was done using a simple Volume

(old) to Volume (new) ratio with the results are shown in Table 5-3:

Table 5-3 Temperature Adjusted Material Densities

Vol Ratio

Old

Density

New

Density

g/cc g/cc

UN 0.967 13.924 13.47

BeO 0.978 2.859 2.80

Nb1Zr 0.971 8.59 8.34

Re 0.973 21 20.43

This ensures that while the volumes of components in the core are changing, the mass of

each material remains constant. Finally, the cross sections of the materials being used in

the core were corrected to those at operating temperature The vast majority of the

materials in the core do not have readily available cross sections at the desired

temperature. The NJoy code was used to generate cross sections for the different

materials in the reactor [MacFarlane, 1994]. The temperatures used were the peak values

from the FEPSIM model. These temperatures are probably somewhat high, but are a

much better approximation than using room temperature cross sections. This run resulted

in a k-effective of 1.022 ±0.001. The net reactivity swing of the system was ~$2. With a

total excess reactivity of $5.7 this leaves another $3.7 for burnup and other losses.

46

5.4 Accident Scenarios:

National expectations for reactors require that any system that would be launched into

orbit should not pose a significant risk to the environment or the public in case of a

launch accident [DOE, 1982]. Generally this requirement is interpreted to mean that the

reactor being designed cannot go critical for any credible accident scenario. Ideally,

under accident conditions, the safety margin would be large enough to allow for things to

go worse than expected and still be safe. The reactor will be examined for three major

accident scenarios: immersion in water, immersion in water with flooding of the core,

and immersion in wet sand with flooding of the core. These three accidents represent the

credible worst case scenarios commonly examined for space reactor proposals. For these

accidents, the design goal is for the reactor to maintain a multiplication factor of less than

0.985.

Increasing the thickness of both the rhenium liner and the uranium nitride pin were

required in the reactor design to meet the safety conditions. During normal operations

the rhenium had a negative effect on the k-effective but was countered by the extra fuel in

the core. In two of the three accident scenarios, the neutron spectrum is more thermal

than during normal operation due to the addition of water to the core. For the dry sand

accident case, the spectrum is faster than the normal operation case. For the accident

scenarios, the extra thermal absorption of neutrons from the additional rhenium

dominated the effects from the additional fissionable fuel.

47

5.4.1 Immersion in Water

In this scenario the reactor is immersed in water, and the gas flow region is flooded. It is

assumed that the radial reflectors are all removed by any splashdown into water. This

scenario results in the moderation of the fast neutrons and normally the increase in cross

section with decreased neutron energy would result in an increase in reactivity. The

inclusion of rhenium in the core between the fuel and the Nb1Zr cladding negates this

effect, as it is a Spectral Shift Absorber. SSA’s are materials that are relatively

transparent to neutrons in the fast spectrum but a massive absorber at the lower end of the

energy spectrum [King, 2005].

Figure 5-4Water Immersion Accident Scenario

The neutron multiplication factor (k-effective) was 0.964 ±0.001, well below the desired

value of 0.985.

5.4.2 Immersion in Wet Sand with Water Flooding

In this scenario the core was immersed in wet sand (70% sand by volume, 1.924 g/cc)

with the core flooded with water. This accident scenario includes a variety of negative

aspects: the water is moderating the neutrons to a significant degree, and the sand is an

Water

No Reflector

48

excellent reflector. This accident scenario has often been a more challenging one than

the water immersion case.

Figure 5-5 Wet Sand, Water Scenario

For this case the k-effective was 0.975 ±0.001; much closer to the margin of 0.985

5.4.3 Dry Sand Burial

This accident scenario involved the reactor being buried in SiO2 with a density of 1.5

g/cc. This is a median point between the density of loose sand (1.4 g/cc) and. dry, packed

sand (1.6 g/cc). The core is not flooded in this case, and the radial reflectors have been

removed. While little moderation occurred in the sand and SiO2 is an inferior scattering

medium to BeO (and is also without the (n,2n) reaction), the reactor was going from

somewhat reflected to essentially infinite reflection conditions. In a variety of cores this

had led to a net increase in reactivity. This effect is highly dependent on the density of

the sand. This core had a k-effective of 0.981 ±0.001 for this scenario.

Wet Sand

Water

49

Figure 5-6 Sand Burial Accident Scenario

The results from these runs can be found in table 5-3

Table 5-4 Accident Scenario Results

Accident

Conditions

K-Effective

Full Water Immersion, Flooded, No

Reflector

0.96387 ±0.001

Full Wet Sand Immersion, Flooded,

No Reflector

0.97484 ±0.001

Full Dry Sand Immersion, No

Reflector

0.98101 ±0.001

These results show that the reactor is substantially subcritical in all of these accident

scenarios. The most problematic of the accident cases is the dry sand immersion case,

which is also the least credible of the accident scenarios. The likelihood of a reactor

coming down intact and burying itself in sand is low. At high velocities even water starts

Dry Sand

50

looking like concrete on impact. None of these scenarios included modeling of

deformation of the reactor core.

5.5 Alternative Cores:

Two major variations on the primary core were analyzed. One core was a test bed

reactor that would use low enrichment fuel of roughly similar dimensions in a thermal

assembly to simulate some of the effects of the reactor while still falling below the

category III threshold. The Category-III threshold is less than 6 kg of 235U at any

enrichment, or it can be less than 50 kg of 235U at an atom enrichment of less than 50%.

This will reduce the security costs relating to the reactor. The other core places a boron

carbide cylinder along the centerline of the reactor to provide an additional safety margin

during the launch of the reactor.

5.5.1 Category-III Reactor

Significant cost savings in security and development can be had if the reactor uses a

Category-III level of SNM rather than Category I. Security for facilities with highly

enriched fuel on site can cost roughly $30 M per year and the slow down in experimental

operations caused by this security can cause an additional cost. The first option requires

a well-moderated reactor; the second requires a somewhat moderated reactor. The first

option may end up being very limited in power and lifetime because of fuel burnup; the

second option has more latitude for power and burnup.

A Category-III space reactor was not designed for this study. But consideration was

made for a Category-III reactor that would simulate the behavior of a Category-I space

reactor. The intent would be to make preliminary nuclear-heated ground tests less

51

expensive. There are drawbacks however. The similarity between the test reactor and

the final design would be somewhat limited. Thus far, the Category III core has not

proven safe for the common launch accident scenarios. The two alternatives that have

been examined are listed below.

One possible design uses pins that are at the upper limits of category III definition

(49%, 49kg 235U) to achieve the desired results. In this case, the size of the fuel pins

would have to be reduced or the number of pins would have to be reduced. Specifics of

this core will be covered in the first alternative core. Another method for achieving the

desired results is to use a lower enrichment fuel and maintain the fuel pin geometry of the

original Category I reactor as much as possible. This design will be detailed in the

second alternative core description. Both of the designs share many similarities;

particularly that the reduced 235U enrichment requires the reactors to be thermal systems,

which limits the usefulness of comparison between the space reactor and the test-bed.

The changeover to a thermal system requires the use of a high temperature moderator.

The primary material being investigated is yttrium hydride. Yttrium hydride appears to

be a stable high-temperature moderator, though data on its materials and neutronics

characteristics is limited. Yttrium Hydride would replace the Nb1Zr matrix between the

fuel pins in the original designs and would likely have a cladding of some material to

minimize hydrogen leakage. In these runs, a generic stainless steel was used to estimate

the neutronic impact of such a liner; a more temperature appropriate choice will have to

be made. To get sufficient moderation the pitch of the reactor will increase significantly

when compared to the original core.

52

5.5.1.1 Alternative 1 Reactor

This section covers the Reactor that uses 49% enriched 235U, 49 kg fuel. The

Alternative 1 reactor uses smaller radius fuel pins enriched to 49% and was the simpler of

the two cores to achieve a supercritical configuration with. The active length of the older

source core was retained, the cross sectional flow are increased, and the overall core

radius decreased. Cross sectional screenshots of the design are shown in Figure 5-7

through Figure 5-9.

Figure 5-7: XZ Plane Section of Reactor Alternative 1

Air

Helium/Xenon Coolant

BeO Reflector

YH2/Fuel Pin Matrix

Hastelloy X PV

53

Figure 5-8 XY Cross Section of Reactor Alternative 1

Figure 5-9 Close up of XY plane, Alternative 1 Reactor

A further parameter search increasing the pitch between the fuel pins and placing extra

yttrium hydride in the matrix was done. The effect of pitch on k-effective is shown in

BeO Reflector

Helium Xenon Coolant

Fuel Pin

YH2 Matrix

YH2 Matrix


UN Fuel

Nb1Zr Clad

Rhenium

SS Liner

Hastelloy X PV

54

Figure 5-10. It shows that the minimum dimensions design was nowhere near the

optimally moderated case and that there was a great deal of margin available for

achieving criticality if needed.

0.96

0.98

1

1.02

1.04

1.06

1.08

1.1

1.12

1.14

1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8

Pitch

K-E

ffect

ive

Normal Operation

Core Flooded, ReflectorsClosed

Core Flooded, WaterImmersion, Reflectors off

451 Pin Type 1 Reactor49% Enriched UN FuelYH2 Matrix

Figure 5-10 K-Effective Vs Pitch, Alternative 1 Core

Some downsides to this design are inherent in the decision to go with a thermal

system. Several of the launch accident scenarios provide an unacceptable risk of the core

going supercritical. Any scenario where water or additional moderator enters the core is

likely to result in a supercritical configuration. This largely was the result of the need to

decrease the thickness of the rhenium layer in the fuel to allow for a critical thermal

system, and thus is probably unavoidable.

55

5.5.1.2 Alternative 2 Reactor

The Alternative 2 reactor is an attempt at designing a lower enrichment reactor

that will fulfill the requirements of a Category III reactor while allowing for a greater

degree of similarity with the source reactor than the above Alternative 1 design. To

ensure this similarity the radial dimensions of the uranium nitride fuel is the same as the

Category 1 reactor source, but the fuel length, enrichment, pitch, and the number of fuel

pins were all varied. When the fuel length was reduced the fuel was replaced with

additional BeO reflector. The gas flow gap was held fixed. This meant that the flow

channels for the gas are the same as those in the original design. Another change that

was necessary was a slight reduction in the thickness of the rhenium layer between the

UN fuel and the Nb1Ze clad from .1mm to .07mm. This reduced the losses in the

rhenium due to thermal neutron absorption. The substitution of yttrium hydride for the

Hastelloy increases the radius of the core significantly. The first core that was critical

employed a combination of the variables listed to achieve criticality: 397 fuel pins, 40cm

fuel length, and 39% enriched fuel. There are probably a variety of ways to achieve the

desired results; this was the first combination that met the requirements. Achieving the

desired results only using 2 of the 3 primary variables (active length and pitch) has

proven difficult. An alternative that retains the original active length and pitch, but

reducing the number of pins and increasing the enrichment might be viable and worth

investigating. Figure 5-11 through Figure 5-123 show a series of cross sectional slices of

the core.

56

Figure 5-11 XZ Cross Section of Alternative 2 Reactor

Figure 5-12 XY Cross Section of Alternative 2 Reactor

Air


BeO Reflector

YH2/Fuel Pin Matrix

YH2 Matrix

Fuel Pin


BeO Reflector

Hastelloy X PV

Hastelloy X PV

57

Figure 5-13 Close up of XY Plane, Alternative 2 Reactor

5.5.2 Boron Carbide Central Safety Rod

The ability to launch a reactor into orbit with a negligible chance of accidental

criticality is one of the keys to having a successful program. Other work has indicated

that dropping the reactor on to an unyielding surface could cause the reactor to become

critical as the fuel pins deformed [Lenard, ]. Some runs for a gas-cooled reactor

indicated that if compression caused a reduction in the pitch of the fuel pins the core

could also become critical. Several ways of mitigating the impact on the system were

investigated. Placing a large rod of polyethylene with a thin layer of boron carbide was a

fairly effective way of mitigating the impact. Figure 5-14 and 5-15 show two slices of

the core that illustrate the differences quite well.

YH2 Matrix


UN Fuel

Nb1Zr Clad

Rhenium

SS Liner

58

Figure 5-14 XY Plane Section of Internal Control Rod Reactor

This XY plane section shows the safety rod inserted in the center of the reactor.

Figure 5-15 XZ Plane Section of Internal Control Rod Reactor

59

Figure 5-15 shows an XZ cross sectional slice of the reactor. The B4C rod displaces 37

fuel pins from the center of the core but 36 of these are added back to the outer ring of

fuel pins with minimal impact on the size of the core. With the rod in place and the

reflectors closed the multiplication factor is 0.993. This is close to the case without the

control rod because of the transparency of materials at high neutron energies. During

water immersion accident scenarios the rod results in a significant decrease in excess

reactivity .940 ±0.002 Without the rod and with the reflectors closed the multiplication

factor is 1.005 ±0.001, well below what was seen in the original design. Additional fuel

would have to be added to make this core practical. These changes complicate the design

significantly. To avoid the problems of welding dissimilar materials the ‘thimble’ for the

control rod has to be made of the same material as the pressure vessel but its placement

along the centerline of the core increases the radiation flux it sees and the temperature at

that location. Neither of these things are good as HastX was a marginal choice in the first

place at the temperatures at outer surface of the cylinder. Either a material that can be

reliably welded to HastX has to be found or the thimble needs to be cooled.

Table 5-5 shows details all of the designs.

Table 5-5 Reactor Dimensions and Compositions

Core Primary Design Alternative 1

Alternative 2 B4C Rod

Gas Properties Coolant He/Xe He/Xe He/Xe He/Xe He fraction 70/30 70/30 70/30 70/30 Total Mass (kg) 1073 910 1789 1070 Mass U235 186 49 49 185

60

Reactor Core, Vessel, Reflectors Type Pin-matrix Pin-matrix Pin-matrix Pin-matrix Reactor Material Properties Fuel material UN UN UN UN Clad material Nb1%Zr Nb1%Zr Nb1%Zr Nb1%Zr Clad liner material Re Re Re Re Wire wrap material Re Re Re Re Moderator material N/A YH2 YH2 N/A Matrix (core block) material Nb1%Zr YH2 YH2 Nb1%Zr Matrix wall material SS304 SS304 N/A Pressure vessel material HastX HastX HastX HastX Radial reflector material BeO BeO BeO BeO Axial reflector material BeO BeO BeO BeO Lower grid material Nb1%Zr HastX HastX Nb1%Zr Upper grid material Nb1%Zr HastX HastX Nb1%Zr Coolant material He/Xe He/Xe He/Xe He/Xe Fissile material U-235 U-235 U-235 U-235 Fuel enrichment 0.9315 0.4899 0.3789 0.9315 Reactor Radial Dimensions Radius of the fuel (UN only) (m) 4.54E-03 3.21E-03 4.44E-03 4.54E-03 Thickness of the fuel gap (m) 6E-05 5E-05 7E-05 6E-05 Thickness of the liner (m) 6.2E-04 5E-05 1E-04 6.2E-04 Thickness of liner-clad gap (m) 4E-05 5E-05 6E-05 4E-05 Thickness of the clad (m) 4.4E-04 1E-04 3E-04 4.4E-04

61

Thickness of coolant channel (m) 1.65E-03 1.64E-03 1.43E-03 1.65E-03 Width to thickness ratio of rectangular wire 1 1 1 1 Pitch of the wire wrap (m) 0.2 0.2 0.2 0.2 Number of wires wraps per pin 2 2 2 2 Thickness of matrix between pins (m) 7.8E-4 1.14E-02 1.3E-02 7.8E-4 Number of fuel pins 451 451 397 450 Thickness added to circle to get matrix radius (m) 0.016 0.016 0.016 0.016 Thickness of matrix insulation (He/Xe) (m) 0.0002 0.0002 0.0002 0.0002 Thickness of the matrix & dome baffle (m) 1E-06 1E-06 1E-06 1E-06 Thickness of the downcomer channel (m) 0.008 0.008 0.008 0.008 Thickness of the pressure vessel (m) 0.003 0.003 0.007 0.003 Thickness of the reflector gap (m) 0.001 0.001 0.001 0.001 Thickness of the radial reflector w/o clad (m) 0.15 0.09 0.16 0.15 Reactor Axial Dimensions Length of the active fuel in the reactor (m) 0.52 0.52 0.40 0.52 Length of the axial fuel plenum (m) 0.03 0.03 0.03 0.03 Length of upper axial reflector (m) 0.05 0.05 0.11 0.05 Length of lower axial reflector (m) 0.05 0.05 0.11 0.05

62

Length of each end cap (m) 0.005 0.005 0.005 0.005 Length of the lower grid plate & pin flow orifices (m) N/A N/A N/A N/A Length of the outlet (upper) plenum (m) 0.005 0.005 0.005 0.005 Length of the inlet (lower )plenum (m) 0.005 0.005 0.005 0.005 Other Axial peak-to-avg ratio in core 1.2 1.2 1.2 1.2 Radial peak-to-avg pin power 1.2 1.2 1.2 1.2 SiO2 fraction in wet sand 0.64 0.64 0.64 0.64 K-Effective of Core

1.037 ±0.001 1.0758 ±0.001 1.004 ±0.001 1.005 ±0.001

5.6 Quality Control Runs

The following results represent an effort to better understand how the reactor would

operate and to increase confidence in the proposed design. The first of these runs ensures

that the neutron population was adequately sampled. A series of runs using a different

seed for the random number generator should generate a distribution of results. If 2/3 of

the runs are within one sigma of the average, then it is likely that the model is being

adequately sampled. If not, there are a variety of possible reasons for the problem. One

possibility is that insufficient particles are being or a significant portion of the core is

being ignored. This can work in both ways: the sampling happens predominantly in the

outer layer of the core, depressing the k-effective or the sampling happens mostly in the

center of the core, underestimating leakage. A total of twenty different runs were done

for the core and the results are shown in Figure 5-16

63

Multiplication Factor Vs Run, different RNG seeds

1.033

1.034

1.035

1.036

1.037

1.038

1.039

1.04

0 5 10 15 20 25

Run #

Mul

tiplic

atio

n

AverageIndividual RunsLower BoundUpper Bound

Figure 5-16 Multiplication vs. Run with different RNG seeds

The runs show that between 6 and 7 of the 20 runs done fall outside the one sigma

deviation, exactly what would be expected. This give an average multiplication factor of

1.037 ±0.001. The second set of runs was focused on what the spectrum of the neutrons

causing fissions in the reactor during accident cases. The results are shown in Figure 5-

17

64

Distribution of Fission Causing Neutrons

0

10

20

30

40

50

60

70

80

90

Normal Wet Sand Water Dry Sand B4C, Wet Sand

Case

%% Thermal (<.625 eV)

% Intermediate (.625 eV – 100 keV)

% Fast (>100 keV)

Figure 5-17 Distribution of Fission Causing Neutrons

During normal operation the vast majority of fissions are caused by fast or intermediate

energy neutrons, which is what one would expect in a slightly moderated core. During

the accident case where water gets into the core, the distribution of the fission causing

neutrons swings heavily towards the intermediate range with the thermal range neutrons

showing a significant increase. The dry sand accident scenario shows the opposite: the

spectrum of the neutrons causing fission actually gets somewhat faster. Finally, the case

with a B4C control rod and wet sand immersion looks similar to the normal wet sand

case, but with a more thermalized distribution.

Another series of runs were done to determine the effect of the Nb1Zr cladding on

the neutronics of the system. One of the reasons for doing these runs was the possibility

that Nb1Zr might end up being incompatible with the UN fuel and there was a desire to

65

know how large a part it played in determining the neutron multiplication factor of the

core. The runs took a single fuel hexagon and put reflective boundaries around it. This

effectively made the core an infinite system. Removing the cladding increased the k-

effective of the system by 0.0056, or 0.86$. This was with the standard deviation being

much smaller than the neutron multiplication factor swing, averaging 0.001. Thus, there

is a sufficient margin that if an alternative cladding to become a necessary, it would not

require massive reworking of the core to incorporate it.

66

6 Thermal Evaluation of Core and Remainder of System

Neutronics effects were not the only aspect of the system to be examined for this

design. A basic evaluation of the properties related to the power conversion system was

also done. The Fission Electric Power SIMulation spreadsheet set was used for thermal

analysis in a variety of ways [Lipinski, 2002]. Its inputs included an approximation of

the reactor core dimensions used in MCNP, dimensions and materials composition for the

remainder of the system (radiator, piping, etc) and inlet temperatures for the turbine.

FEPSIM takes this information and estimates, among other things, the peak fuel, liner,

cladding, and pressure vessel temperatures. This allowed the expanded dimensions of the

different components at operating temperature to be determined and used in later MCNP

runs. The efficiency of the power conversion system and whether the pressure losses

were severe enough to cause the Brayton Cycle to stall were also generated.

Figure 6-1 FEPSIM components with State temperatures

Com-pressor

Alternator

Reactor: Pin geo-metry

T4: 1144K T1: 350-470k

T3: 850k

T2: 500K T5: 900K

T6: 500K

Turbine

Recuperator

Bypass flow

CPL or LHP Conical or Flat Rectangular Radiator

kWe DC

Radiator Heat Exchanger

Shield

Closed Brayton Cycle

67

6.1 Radiator

The radiator used in the reactor has a drastic effect on the output power of the reactor.

For the models used, the temperature of the gas coming out the reactor is held at a fixed

temperature. Gas passing through the turbine has a certain temperature drop. The gas

then reaches the radiator which further cools the gas. The amount that the gas is cooled

in the radiator is highly dependent on the size of the radiator. This becomes extremely

important when determining the amount of work the compressor has to do. At lower

temperatures and pressures, the work done by the compressor to compress the gas to the

reactor inlet conditions is low. As the temperature increases the amount of work

increases, decreasing the net power output.

Initially, the radiator dimensions were carried over from a higher power reactor

that had 200 m^2 of radiating area. This was an enormously oversized radiator that

resulted in the inlet temperature for the reactor being low and the power conversion

efficiency high (as high as 37%). To see what effect radiator parameters could have on

the core, 4 different sizes of radiator at 5 different thermal power levels were examined.

The radiator areas were, from low to high, 80 m^2, 98 m^2, 140 m^2, and 200 m^2. The

thermal power levels examined were 200 kW, 250 kW, 300 kW, 350 kW, and 400 kW.

Figure

6-2 shows the electric power vs. thermal power for the different radiators.

68

30

40

50

60

70

80

90

100

110

150 200 250 300 350 400 450kW Thermal

kW E

lect

ric 80 m^298 m^2140 m^2200 m^2

Figure 6-2 Electric Output vs. Thermal Output

This plot shows that for the smaller radiators the power output peaks in the range under

examination and then starts to drop off. As the thermal power goes up for the designs

with the larger radiators the electric power keeps on rising. If these plots were extended

to even higher thermal power outputs they would peak also. This is a result of the

different rates at which the electric conversion cycle efficiency is changing. This is

shown on Figure 6-3

69

5%

10%

15%

20%

25%

30%

35%

40%

150 200 250 300 350 400 450kW Thermal

Con

vers

ion

Effic

ienc

y

80 m^298 m^2140 m^2200 m^2

Figure 6-3 Conversion Efficiency Vs Thermal Output

The thermal efficiency of all the systems are falling, but at different rates. The smallest

radiator drops from 27% efficiency at 200 kWth to 11% efficiency at 400 kWth. This

means that in the process of doubling the thermal output of the reactor the conversion

efficiency has dropped by more than half, causing a net loss in power output. On the

other hand, the 200 m^2 radiator drops from an efficiency of 36% to 27%, a drop of a

third, while the thermal power doubles, generating a net increase in power output

(400*.27 > 200*.36). Finally, the weight of the system needs to be examined. One way

of expressing this is the inverse specific power (expressed in estimated system kg /power

output) vs. the thermal power of the system. This is shown in 6-4. The density per unit

surface area of the radiator was fixed at 5 kg/m2

70

30

40

50

60

70

80

90

150 200 250 300 350 400 450kW Thermal

Spec

ific

Pow

er (k

g/kW

e)

80 m^298 m^2140 m^2200 m^2

Figure 6-4 Specific Power Vs Thermal Output

This implies that larger radiators allow for more mass effective production of power. The

only variables are the thermal output and the size (and mass) of the radiator. For the 80

m2 radiator as the thermal power goes up the efficiency goes down while the weight of

the radiator stays the same, so the effectiveness of the system drops. The general result of

this is that the size of the radiator one chooses for the reactor depends on several

considerations. What is the desired power output? How much margin is to be included to

mitigate the effect of system degradation? Is the system intended to be a variable power

system operating at a variety of levels or a system that has constant power output? All of

these questions need to be answered. All of these Figures show that if the desired output

is in the 50-60 kWe range the radiator only needs to be 80-98 m^2. If higher outputs are

desired (greater than 100 kWe range) a larger radiator of about 200m^2 would be needed.

71

For these designs, the primary fixed property was the turbine inlet temperature.

Variables such as the gas flow rate and the pressure drop was variable. This means that

the temperature of the fuel is also variable, so one has to examine that as well to ensure

that safety margins have not been violated. These runs show that the power output is

more dependent on the size of the radiator than the size of the reactor. Significant

increases in power for a given core can be achieved by increasing the size of the radiator..

72

7 Conclusions and Future Work

7.1 Conclusions

The primary goal of this thesis was to do design work on a GCR-CBC reactor. The

focus was on the neutronics of the core with basic analysis of the other components of the

design. There is a significant region where a GCR pin type core looks to be feasible.

While materials issues warrant examination, no critical problems were found. A variety

of MCNP runs were done covering a wide range of situations and none of them cause a

problem that a small amount of tuning of the design could not resolve. The thermal

radiation component showed just how complicated a reactor system can be. The tradeoffs

between radiator size, electrical output, and thermal efficiency are extremely

complicated.

The reactor was to meet some specific requirements. The design was to be a gas-

cooled, pin type reactor. This was a fairly easy requirement to meet. The desired power

output was100 kWe. This was more complicated, depending on the efficiencies of the

Brayton cycle, the thermal output of the reactor and all the related conditions to ensure

that the Brayton cycle was stable. The requirement that the reactor have a 10 year

lifespan at full power was easily met given that the fuel was highly enriched. For a

thermal neutron spectrum system the lifespan requirement might be more of an issue.

Finally, there was to be some investigation into how the reactor would be modified to

make it work in a Martian atmosphere. This requirement ended up being a fairly

problematic in that it required a great deal of materials engineering knowledge.

For GCR-CBC reactor there is a great deal of interaction between the components

and a system wide view is important. It is insufficient for the reactor to be critical: if the

73

gas flow across the pins generates enough pressure loss to stall the Brayton cycle the

system will not work. The power output is highly dependent on the amount of waste heat

the radiator can vent. Design of a functional system will require a great deal of

interdiscipline collaboration.

7.2 Future Work

There is a great deal of potential work to be done on the reactor design. Materials

engineering related problems are one area with a great deal of potential work. Expanding

on the limited data on HastX and Nb1Zr compatability is one area that needs work. The

amount of oxygen and other contaminants that would be leeched from the Hast-X and

how much of these contaminants would diffuse into the Nb1Zr components of the reactor

is one of the potential problems. Further testing of HastX in a mars-like atmosphere to

characterize its interaction with that environment would also be valuable. Further data on

how rhenium and UN interact and what occurs when they are both irradiated are also

potential topics. There were some more conventional mechanical engineering problems

that were unaddressed. While the strengths of the materials used was examined, the

stresses that would be generated were not examined. The pressures generated by the

release of fission gasses are one possible problem. Another is stresses generated by the

different expansion rates of the materials. Shielding for radiation was not addressed in

this thesis. Finally, the design of the Brayton cycle and any optimization that would be

related to that part of the reactor was not addressed in this thesis.

74

References

Angelo, J. A., Jr., and D. Buden, Space Nuclear Power, Orbit Book Company, Inc., Malabar, FA, 1985. Biaglow, James A., Rhenium Material Properties, ASME, SAE, and ASEE, Joint Propulsion Conference and Exhibit, 31st, San Diego, CA, AIAA-1995-2398 July 10-12, 1995. Brown, Nicholas, Gas-Cooled Reactor Vessel Design and Optimization with COSMOSFloWorks, American Nuclear Society Student Conference, Rensselaer Polytechnic Institute, NY, March 31, 2006. Brown, W.F., et. al., Aerospace Structural Metals Handbook, CINDAS/Purdue University, code 4112, 1992 ed. Davis, J.E., Design and Fabrication of the Brayton Rotating Unit, NASA/CR-1870, March 1972. DiStefano, J. R., Hendricks, J. W., Oxidation of Nb-1Zr in Space System Environments, ORNL-TM-11423, Oak Ridge National Laboratory, 1990. El-Wakil, M. M., Powerplant Technology, pp. 309-351 433, 1984. Furlong, Richard R. U.S. Space missions using radioisotope power systems, Nuclear News, April 1999, page 28. Horak, J. A., Data Correlations and Creep Equations for Nb-1%Zr and PWC-11, ORNL Letter 0514/20/85. May 14, 1985. Johnson, C. E., Thermophysical and Mechanical Properties of Advanced Carbide and Nitride Fuels., ANL-AFP-26., June 1976. Keiffer, H. H., Jakosky, B. M., Snyder, C. W., and Matthes, M. S. (Editors), Mars, Univ. of Arizona Press, Tuscon, AZ, 1992. King, J.C., El-Genk, M. S., Spectral Shift Absorbers for Fast Spectrum Space Nuclear Reactors., in proceedings of Space Technology and Applications International Forum (STAIF 2005), edited by M. El-Genk, AIP Conference Proceedings 746, New York, 2005. Lipinski R. J., FEPSIM, Sandia Internal Space Reactor Power System Design Code, Unpublished, 2002.

75

MacFarlane, R.E., Muir, D. W., The NJOY Nuclear Data Processing System Version 91, LA-12740-M, October 1994.

Niobium Metallurgical Products, http://www.dhcommunications.com/success/shared/niobium%20brochure.pdf, accessed March 15, 2006. NIST Chemistry WebBook http://webbook.nist.gov/chemistry/ accessed March 1st, 2006. Paxton D.M., and B.J.Makenas, “Postirradiation Examination Report for the SP-3RR Test, WHC-SP-1053, 1993. Selection of Power Plant Elements for Future Reactor Space Electric Power Systems Los Alamos National Laboratory, LA-7858, 1979. Staff of the Solar Dynamic Power Systems Branch, Solar Dynamic Power System Development for Space Station Freedom, NASA/RP-1310, July 1993. Summary of Design and Analysis of the SNAP-50/SPUR Reactor Coolant Pump. PWAC-467. August 10, 1965. Tagawa, H., Phase Relations and Thermodynamic Properties of the Uranium Ntride System, Journal of Nuclear Materials, pp. 78-79, Vol. 51, 1974. Touloukian, Y. S., et al, Vol. 1 Thermal Conductivity Metallic Elements and Alloys, Thermophysical Properties Research Center, Purdue University. IFI/Plenum, New York, 1970. Touloukian, Y. S., et al, Vol. 12 Thermal Conductivity Metallic Elements and Alloys, Thermophysical Properties Research Center, Purdue University. IFI/Plenum, New York, 1975. Touloukian, Y. S., et al, Thermophysical Properties of Matter V. 13., Thermophysical Properties Research Center, Purdue University. IFI/Plenum, New York, 1979. P. 1153. Nuclear Safety Criteria and Specifications for Space Nuclear Reactors, Dept. of energy, Document OSNP-1, Rev. 0, Washington, D.C. Dated Aug 1982, Watson, W., Short Time Properties of Advanced Alloys Tested at CANEL., Pratt and Whitney Aircraft- CANEL., April 20, 1965. X-5 Monte Carlo Team, MCNP – A General Monte Carlo N-Particle Transport Code, Version 5., Los Alamos National Laboratory, LA-UR-03-1987, April 2003.

76

Distribution List external recipients names and addresses 1 MS0736 John Kelly 6870 1 MS0736 Ronald J. Lipinski 6872 4 MS1136 Curtis D. Peters 6872 3 MS1136 Paul Pickard 6872 1 MS1146 Steven A. Wright 6872 1 MS1146 Roger X. Lenard 6872 2 MS9018 Central Technical Files 8944 2 MS0899 Technical Library 4536 For LDRD reports, add: 1 MS0123 D. Chavez, LDRD Office 1011 For CRADA reports add: 1 MS0115 CRADA Administration 10112

A 50-100 kWe Gas-cooled Reactor For Use On...

Documents

Transcript of A 50-100 kWe Gas-cooled Reactor For Use On...